Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Apr 1.
Published in final edited form as: Child Abuse Negl. 2020 Feb 13;102:104361. doi: 10.1016/j.chiabu.2020.104361

Child Maltreatment and Depression: A Meta-Analysis of Studies Using the Childhood Trauma Questionnaire

Kathryn L Humphreys 1,*, Joelle LeMoult 2,*, John G Wear 3, Hannah A Piersiak 1, aaron Lee 4, Ian H Gotlib 5
PMCID: PMC7081433  NIHMSID: NIHMS1560768  PMID: 32062423

Abstract

Background

Researchers have documented that child maltreatment is associated with adverse long-term consequences for mental health, including increased risk for depression. Attempts to conduct meta-analyses of the association between different forms of child maltreatment and depressive symptomatology in adulthood, however, have been limited by the wide range of definitions of child maltreatment in the literature.

Objective

We sought to meta-analyze a single, widely-used dimensional measure of child maltreatment, the Childhood Trauma Questionnaire, with respect to depression diagnosis and symptom scores. Participants and Setting: 192 unique samples consisting of 68,830 individuals.

Methods

We explored the association between total scores and scores from specific forms of child maltreatment (i.e., emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect) and depression using a random-effects meta-analysis.

Results

We found that higher child maltreatment scores were associated with a diagnosis of depression (g=1.07; 95% CI, 0.95−1.19) and with higher depression symptom scores (Z=.35; 95% CI, .32−.38). Moreover, although each type of child maltreatment was positively associated with depression diagnosis and scores, there was variability in the size of the effects, with emotional abuse and emotional neglect demonstrating the strongest associations.

Conclusions

These analyses provide important evidence of the link between child maltreatment and depression, and highlight the particularly larger association with emotional maltreatment in childhood.

Keywords: child maltreatment, depression, abuse, neglect, meta-analysis

Depression is a significant public health concern; indeed, major depressive disorder (MDD) is the leading cause of disability worldwide (World Health Organization, 2017). Understanding the etiology of depression, and in particular mutable factors that may play a causal role, is critical for reducing risk for this recurrent and debilitating disorder (Liu, 2017). Prospective studies have documented that greater adversity in childhood is associated with more chronic depression (Klein & Kotov, 2016), more severe depression (Rhebergen et al., 2012), and a longer time to remission (Fuller-Thomson, Battiston, Gadalla, & Brennenstuhl, 2014). The role of early adversity in increasing risk for the subsequent development of depression is substantial; in fact, Kessler et al. (2010) estimated that almost 25 percent of population-attributable risk is due to early adversity.

Among early adverse experiences, child maltreatment is a particularly potent risk factor for depression. Previous meta-analyses examining child maltreatment and depression have found that experiencing any form of maltreatment (treated statistically as the presence or absence of maltreatment) was associated with more than a two-fold increase in risk for depression in adulthood (Li, D’Arcy, & Meng, 2016), and with the development of chronic, or recurrent, depression (Nanni, Uher, & Danese, 2012). Although sexual abuse has received the most empirical attention (see Liu, 2017), it is noteworthy that different types of maltreatment frequently co-occur (Petersen, Joseph, & Feit, 2014). Thus, rather than focus on a single type of maltreatment, it is important to characterize the relation between different types of child maltreatment and depression. This perspective is supported by the emerging theory that early experiences that are characterized by threat (e.g., abuse) have different effects on the emergence of psychopathology than do early experiences characterized by a lack of species-expected input (e.g., neglect; Humphreys & Zeanah, 2015). Further, although physical, sexual, and emotional abuse have all been linked to depression (Mullen, Martin, Anderson, Romans, & Herbison, 1996), their different prevalence rates (Edwards, Holden, Felitti, & Anda, 2003), and their differential links to depressogenic features (e.g., low self-esteem following emotional abuse; Mullen et al., 1996), underscores the importance of careful examination of different forms of maltreatment with depression.

Previous meta-analyses examining the association between maltreatment and depression are informative. However, while important, all are limited either by small number of available studies (e.g., 8 for Li et al., 2016; 12 for Infurna et al., 2016; 16 for Nanni et al., 2012) or by considerable variability in how child maltreatment was operationalized (e.g., Norman et al., 2012), which limits comparisons across studies. Given that different definitions, informants, and thresholds for characterizing maltreatment are likely to result in different patterns of findings, there is value in prioritizing the meta-analysis of studies that use a common measure to assess maltreatment. In one such example, Infurna and colleagues (2016) conducted a meta-analysis of studies using the Childhood Experience of Care and Abuse interview (CECA; Bifulco, Brown, & Harris, 1994). They also restricted their inclusion criteria to studies that required a clinical diagnosis of depression. This increased confidence in their conclusions has a trade-off, which is that only 12 studies met inclusion criteria, which limited their ability to conduct moderator analyses. Moreover, experiences of maltreatment, as well as characterization of depression, may better be considered along a dimension (i.e., people vary in the severity of their maltreatment experiences [Humphreys & Zeanah, 2015; King, Humphreys, & Gotlib, 2019; McLaughlin, Sheridan, & Lambert, 2014] and depression to be represented both dimensionally and categorically [Ruscio & Ruscio, 2000]).

Thus, we sought to meta-analyze studies that assessed maltreatment experiences on a continuous scale using the Childhood Trauma Questionnaire (CTQ; Bernstein et al., 2003; Bernstein & Fink, 1998). The CTQ is the most widely used measure of this construct; it has been shown to have acceptable internal consistency, test-retest reliability, and strong convergence with interviews that assess child trauma (Bernstein et al., 1994). The CTQ assesses five types of maltreatment experiences (i.e., emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect) using a Likert-scale approach to assess the severity of each type of experience. By assessing maltreatment using a dimensional approach, and by using a single assessment measure (i.e., the CTQ), our meta-analysis maximizes consistency in the measurement of child maltreatment and increases confidence in the effect size estimates in relation to depression; moreover, this meta-analysis includes the largest set of studies and number of unique participants assessed using a single measure examined to date. Further, unlike prior meta-analyses that vary in the forms of maltreatment that were considered in their assessments, our approach allows us not only to probe associations between depression and overall maltreatment, but also to assess specific types of maltreatment measured at the same time using the same scale. Such an approach will yield insight into whether models indicating that the type of maltreatment or deviation from an expectable environment are differentially associated with depression (including neglect versus abuse; see Humphreys & Zeanah, 2015; McLaughlin & Sheridan, 2016; McLaughlin et al., 2014; or emotional maltreatment versus physical or sexual maltreatment). Finally, by including studies that examined depression using either a group-based approach (e.g., diagnoses) or a continuous approach (e.g., depression symptom scores), we can examine the strength and specificity using two widely used approaches to the assessment of depression.

Method

Study Selection

Each study satisfied the following inclusion criteria: (a) dimensional measurement of child maltreatment using the CTQ (either the long or short form); (b) dichotomous or dimensional assessment of depression; and (c) available data to calculate effect sizes (i.e., standardized mean difference in studies examining depression group and Z in studies examining depression scores).

Search Procedure

We used several strategies, outlined in the PRISMA flowchart (Figure 1), to identify the 190 journal articles with 192 independent samples that were ultimately included in this meta-analysis. First, we conducted computer-based searches using PubMed and Ovid for the following terms (or stems when appropriate) appearing anywhere in the manuscript: (depress* OR MDD) AND (ctq OR “child trauma questionnaire” OR “childhood trauma questionnaire”). Second, we reviewed the bibliographies for additional studies using forward and backward searching. Third, we sent emails describing our meta-analysis and its inclusion criteria to professional membership LISTSERVs of research organizations including the Society for a Science of Clinical Psychology, the Association for Behavioral and Cognitive Therapies, and Division 53. The majority of reviewed studies were excluded due to the presence of confounding medical conditions (e.g., heart disease, diabetes, cancer), the lack of examination of CTQ as a predictor for MDD, and insufficient data for our quantitative analysis.

Figure 1.

Figure 1.

Open in a new tab

Identification of independent studies for inclusion in meta-analysis (PRISMA)

Data Extraction

Two trained raters independently coded each study. When raters provided contradictory judgments, disagreements were discussed and the lead authors made a final determination.

Moderator Variables

When heterogeneous effect sizes were detected, we tested whether potentially important demographic and methodological factors moderated the association between child maltreatment and depression. These moderators were selected on the basis of both recommendations from experts in meta-analysis (Lipsey & Wilson, 2001) and of prior work by the authors (Humphreys, Eng, & Lee, 2013; LeMoult et al., 2019; Muscatell, Humphreys, & Brosso, 2018). We coded the following demographic characteristics: (a) mean age of the sample at the depression assessment (in years) and whether the mean age was above or below age 18 years; (b) sex composition (percent male); and (c) racial diversity (percent Caucasian). We coded the following methodological characteristics of each study: (a) sample size; (b) year published; (c) sample source (i.e., clinic-referred; community; population-based; other); (d) assessment used to determine depression (i.e., Structured Clinical Interview for the Diagnostic and Statistical Manual [SCID] vs. other) or symptom scale (i.e., Beck Depression Inventory [BDI]; Center for Epidemiologic Studies Depression Scale [CES-D] vs. other); (e) whether the original 53 item version of the CTQ was used (vs. 25/28 item short-form); and (f) the language in which the measures were given (i.e., English vs. other).

Calculation of Effect Size

We calculated two different types of effect sizes depending on whether depression was operationalized as a dichotomous (diagnosis of depression) or a dimensional (depression scores) measure. When depression was operationalized as a dichotomous measure, we calculated the Hedge’s g standardized mean difference (SMD) in order to estimate the effect size of the association between child maltreatment total scores and the onset of a diagnosis of depression. An estimate of 0 for the SMD effect size indicated that child maltreatment scores were equivalent in individuals with and without depression, whereas an SMD greater than 0 indicated that the depressed group had higher scores on the CTQ than did individuals without depression, and an SMD less than 0 indicated that the depressed group had lower scores on the CTQ than did those without depression. When depression was operationalized dimensionally, we calculated the bivariate association between child maltreatment scores and depression scores by converting correlations and standardized β to Z values. A Z estimate of 0 indicated no association between child maltreatment and depression, whereas a Z value greater than 0 or less than 0 indicated that maltreatment had a positive or negative association, respectively, with depression scores. The 95 % confidence interval (CI) for the effect size represents the relative precision of the measurement (wider ranges are less precise). For each study, we calculated as many as 12 effect sizes: the two forms of depression measurement (diagnosis and scores) by CTQ total scores and the five types of child maltreatment. These procedures produced 609 total effect sizes estimated from 190 eligible studies. The number of studies were 39 for CTQ total score by depression group and 70 for CTQ total score by depression scores (see Tables 2 and 3).

Table 2.

Summary of Meta-Analysis Statistics by Correlations Between Continuous Depression Scores and Childhood Trauma Questionnaire Scores

Outcome k Coeff. (95% CI) Effect estimate differed from 0 Test for heterogeneity I2 Pooled Z range using leave-one-out analyses Moderators with significant associations
Total CTQ scores 70 Z = .35 (.32–.38) Z = 21.21, p < .001 Q = 418.26, p < .001 84% .35–.36 + in community samples; + using CES-D
Emotional Abuse 81 Z = .38 (.34–.41) Z = 22.15, p < .001 Q = 607.42, p < .001 87% .37–.38 + year published; − in population-based samples
Physical Abuse 66 Z = .22 (.18–.25) Z = 12.70, p < .001 Q = 393.85, p < .001 84% .21–.22 --
Sexual Abuse 72 Z = .20 (17–.23) Z = 14.22, p < .001 Q = 287.82, p < .001 74% .19–.20 --
Emotional Neglect 58 Z = .30 (.26–.34) Z = 15.83, p < .001 Q = 444.84, p < .001 87% .30–.30 − in population-based samples
Physical Neglect 48 Z = .23 (.20–.27) Z = 13.15, p < .001 Q = 237.41, p < .001 80% .23–.24 − in population-based samples
Open in a new tab

Note. CTQ = childhood trauma questionnaire.

Table 3.

Summary of Meta-Analysis Statistics by Depression Diagnosis and Childhood Trauma Questionnaire Scores

Outcome k Coeff. (95% CI) Effect estimate differed from 0 Test for heterogeneity I2 Pooled g range using leave-one-out analyses Moderators with significant associations
Total CTQ scores 39 g = 1.07 (0.95–1.19) Z = 16.98, p < .001 Q = 248.65, p < .001 85% 1.02–1.09 + CTQ-53; + English language
Emotional Abuse 35 g = 0.85 (0.77–0.94) Z = 18.16, p < .001 Q = 64.94, p < .001 48% 0.84–0.87 --
Physical Abuse 35 g = 0.47 (0.37–0.57) Z = 9.42, p < .001 Q = 77.88, p < .001 56% 0.45–0.49 --
Sexual Abuse 35 g = 0.44 (0.36–0.53) Z = 10.17, p < .001 Q = 58.89, p = .005 42% 0.42–0.46 + English language
Emotional Neglect 35 g = 0.96 (0.85–1.08) Z = 16.52, p < .001 Q = 102.43, p < .001 67% 0.93–0.98 + English language
Physical Neglect 35 g = 0.65 (0.53–0.78) Z = 10.39, p < .001 Q = 129.67, p < .001 74% 0.58–0.68 --
Open in a new tab

Note. CTQ = childhood trauma questionnaire. CTQ-53 = original 53 item version of the CTQ (vs. the short form with 25 scorable items).

Statistical analysis

We conducted random-effects models and estimated heterogeneity of effect sizes using the standard Cochran’s Q Test, which indicates the degree of consistency of findings across studies and approximates a chi-square distribution with k–1 degrees of freedom, where k is the number of effect sizes (Hedges & Olkin, 1983). A nonsignificant Q test statistic suggests that the pooled OR represents a unitary effect. When the p-value associated with the Q statistic was equal to or less than .05, we conducted random-effects meta-regression analyses to determine whether the study characteristics described above could explain variability across studies. We assessed publication bias via Egger’s test (Egger et al., 1997). When we observed heterogeneous effect sizes, we conducted leave-one-out sensitivity analyses to test whether a single study unduly influenced effect size estimates. In addition, we examined whether any of the moderator variables predicted significant variance in the effect sizes that had significant heterogeneity. We used STATA 14 to conduct the analyses.

Results

Table 1 presents descriptive information for each included study, including details of demographic and methodological moderators coded and outcomes obtained. Extracted and coded data is available and can be obtained by emailing the lead author.

Table 1.

Study list and features

Study Sample % Male % White Sample Source Age of Depression Assessment Depression Measure CTQ version Language
Aguilera et al., 2009 521 45 NS Community/volunteer 22.9 SCL-90-R CTQ-SF Spanish
Akbaba Turkoglu et al., 2015 120 0 NS Any clinic referred 33.38 BDI CTQ-SF Turkish
Allen et al., 1998 142 0 NS Any clinic referred 37.3 BSI CTQ-53 English
Ammerman et al., 2013 208 0 80 Other 21.27 BDI-II CTQ-SF English
Arata et al., 2005 383 30 71 Community/volunteer 20.4 CES-D CTQ-SF English
Arslan et al., 2015 320 34 NS Community/volunteer 24.62 BSI CTQ-SF Turkish
Auslander et al., 2016 237 0 25 Other 14.9 CDI CTQ-SF English
Aversa et al., 2014 249 100 77 Any clinic referred 29 HAMD CTQ-SF English
Bailer et al., 2014a 162 41 NS Any clinic referred 42.9 SCID-I, PHQ-9 CTQ-SF German
Balsam et al., 2010 669 38 78 Community/volunteer 36.5 CES-D CTQ-SF English
Banducci et al., 2014a 222 56 51 Community/volunteer 11.02 RCADS CTQ-SF English
Banducci et al., 2014b 280 70 NS Any clinic referred 43.3 HAMD CTQ-SF English
Banou et al., 2009 64 0 86 Other 53.4 CES-D NS English
Basu et al., 2013 88 0 52 Community/volunteer 27 SCID-I CTQ-SF English
Bauriedl-Schmidt et al., 2017 81 52 NS Any clinic referred 45.53 NS CTQ-SF German
Bermingham et al., 2012 88 38 NS Any clinic referred 38.77 SCID NS English
Bernet & Stein, 1999 88 50 74 Community/volunteer 42.17 SCID-I, HRSD CTQ-53 English
Blain et al., 2012 182 100 59 Community/volunteer 35.99 BDI-II CTQ-SF English
Blom et al., 2017 26 27 46 Community/volunteer 15.6 RADS-2 CTQ-SF English
Boecking & Barnhofer, 2014 40 40 70 Any clinic referred 36.63 BDI-II, MDI CTQ-SF English
Brown et al., 2016 339 51 72 Community/volunteer 19.00 SMFQ CTQ-SF English
Bruwer et al., 2008 502 41 31 Community/volunteer 16.22 BDI CTQ-SF English
Burns, 2012 996 0 80 Community/volunteer 18.98 BDI-II CTQ-SF English
Caceda et al., 2014 89 42 NS Any clinic referred 34.84 NS NS English
Caldwell et al., 2011 76 0 51 Community/volunteer 28 SCL-90-R CTQ-SF English
Carballedo et al., 2013 133 38 NS Any clinic referred 40.0 SCID NS NS
Carew et al., 2013 47 0 NS Any clinic referred 21.4 BDI-II, HAM-D, MINI NS NS
Carpenter et al., 2009 68 41 NS Community/volunteer 40.12 SCID-I/P CTQ-SF English
Chaney et al., 2014 83 41 NS Any clinic referred 38.22 Prior diagnosis CTQ-SF English
Chen et al., 2017 1705 62 NS Other NS BDI-II CTQ-SF Chinese
Choi et al., 2015 84 0 NS Any clinic referred NS EPDS CTQ-SF Afrikaans, English, Xhosa
Choi et al., 2017 150 0 NS Any clinic referred 25 EPDS CTQ-SF NS
Cisler et al., 2013 38 0 35 Community/volunteer 28.88 SCID-I NS English
Cohen et al., 2017 580 42 29 Community/volunteer 18.25 CES-D CTQ-SF English
Cort et al., 2011 104 0 33 Any clinic referred 31.29 BDI-II CTQ-53 English
Crow et al., 2014 3902 31 NS Other 39.34 BDI-II CTQ-SF English
Cyranowski et al., 2012 335 0 55 Population-based/epidemiological 46.2 SCID-I CTQ-SF English
Dackis et al., 2012 236 0 34 Other 33.8 BDI-II CTQ-SF English
Dannehl et al., 2017 131 36 NS Any clinic referred 36.47 SCID-I CTQ-SF German
Day et al., 2013 112 61 22 Other 16.8 CES-D CTQ-SF English
Ding et al., 2017 6406 52 NS Population-based/epidemiological 12.55 CES-D CTQ-SF Chinese
Douglas & Porter, 2012 105 37 NS Any clinic referred 38.77 Prior diagnosis CTQ-SF English
Du et al., 2016 34 38 NS Other 36.65 Prior diagnosis NS Chinese
Dunlop et al., 2015a 191 0 85 Other 44.2 IDS-SR CTQ-SF English
Dunlop et al., 2015b 140 100 85 Other 44.2 IDS-SR CTQ-SF English
Engelmann et al., 2013 36 25 58 Any clinic referred 37.04 SCID-I, HAMD NS English
England-Mason et al., 2017 140 100 87 Other 32.3 EPDS CTQ-SF English
Fernando et al., 2012 74 36 NS Any clinic referred 33.19 SCID-I NS German
Fernando et al., 2014 111 40 NS Any clinic referred 32.18 SCID-I NS German
Franzke et al., 2015 87 0 100 Any clinic referred 41.32 BDI CTQ-SF NS
Frodl et al., 2017 3036 47 NS Any clinic referred 41.32 Prior diagnosis CTQ-SF NS
Gavin et al., 2011 132 50 47 Community/volunteer 27 DIS CTQ-SF English
Gerke et al., 2006 417 0 58 Community/volunteer 19.9 CES-D CTQ-SF English
Gibb & Abela, 2008 105 49 84 Community/volunteer 9.82 CDI CTQ-SF French
Goldstein et al., 2012 202 46 30 Other 15.93 BSI CTQ-SF English
Goldstein et al., 2013 93 24 16 Other 19.46 CES-D CTQ-SF English
Gradin et al., 2016 50 32 NS Community/volunteer 25.46 BDI, MINI NS English
Grant et al., 2014 39 46 NS Any clinic referred 32.89 SCID-I, HAMD CTQ-SF English
Grassi-Oliveira et al., 2008 49 0 NS Any clinic referred 38.53 SCID-I CTQ-SF Portuguese
Grassi-Oliveria et al., 2009 49 0 NS Any clinic referred 38.49 SCID-I CTQ-SF Portuguese
Grassi-Oliveria et al., 2011 42 0 NS Any clinic referred 39.35 SCID-I CTQ-SF Portuguese
Gratz et al., 2011 225 55 50 Community/volunteer 12.15 RCADS CTQ-SF English
Grosse et al., 2016 394 41 NS Any clinic referred 38.72 MINI CTQ-SF NS
Güleç et al., 2013 150 29 NS Any clinic referred 39.33 SCID, HDRS CTQ-SF Turkish
Hamilton et al., 2016 410 47 49 Community/volunteer 12.84 CDI CTQ-SF English
Harding et al., 2012 157 0 58 Community/volunteer 19.22 BDI-II NS English
Heckman & Westefeld, 2006 138 17 96 Any clinic referred 39.72 TSI CTQ-SF English
Hentze et al., 2016 25 36 NS Any clinic referred 41.52 MADRS CTQ-SF German
Hopwood et al., 2011 (female) 82 0 78 Any clinic referred 15.90 BDI CTQ-SF English
Hopwood et al., 2011 (male) 66 100 78 Any clinic referred 15.90 BDI CTQ-SF English
Hostinar et al., 2017 314 44 63 Population-based/epidemiological 55.3 CES-D CTQ-SF English
Huh et al., 2017 585 46 NS Any clinic referred 36.94 BDI CTQ-SF Korean
Hund & Espelage, 2005 608 0 69 Community/volunteer 20.3 CES-D CTQ-SF English
Hund & Espelage, 2006 608 0 69 Community/volunteer 20.3 CES-D CTQ-SF English
James et al., 2012 286 100 76 Community/volunteer 44.28 HAMD CTQ-SF English
Jessar et al., 2017 204 46 48 Community/volunteer 12.85 CDI CTQ-SF English
Jin et al., 2014 134 100 NS Community/volunteer 45.6 Prior diagnosis NS Malayalam
Jobst et al., 2015 38 68 NS Any clinic referred 46.19 SCID-I CTQ-SF German
Jonas et al., 2013a 280 0 80 Any clinic referred 30.00 CES-D CTQ-SF NS
Jonas et al., 2013b 151 0 76 Any clinic referred 29.05 CES-D CTQ-SF NS
Jovanovic et al., 2010 106 38 NS Other 44.56 BDI, SCID-P CTQ-SF NS
Kecojevic et al., 2015 191 100 64 Community/volunteer 23.7 BSI CTQ-SF English
Khan, 2017 146 0 77 Community/volunteer 32.08 BDI-II CTQ-SF English
Kilimnik & Meston, 2016 222 0 68 Community/volunteer 33.10 BDI CTQ-53 English
Kim et al., 2017 207 41 NS Community/volunteer 27.86 BDI CTQ-SF Korean
Kimbrel et al., 2015 155 93 66 Community/volunteer 40 PDSQ CTQ-SF English
Kimonis et al., 2017 232 100 42 Other 16.75 CES-D CTQ-SF English
Klein et al., 2008 250 0 12 Community/volunteer 35.3 DASS CTQ-53 English
Klein et al., 2009 808 45 86 Any clinic referred 43.6 HAMD CTQ-SF English
Klein et al., 2016 45 27 NS Any clinic referred 42.47 NS CTQ-SF German
Klein, 2014 332 100 74 Population-based/epidemiological 43.7 CES-D NS English
Kounou et al., 2013 181 34 NS Any clinic referred 28.98 Prior diagnosis CTQ-SF French
Krastins et al., 2014 411 24 86 Community/volunteer 29.75 DASS CTQ-SF English
Lang et al., 2004 72 0 56 Community/volunteer 32.73 CES-D CTQ-SF English
Lang et al., 2006 44 0 61 Community/volunteer 29.3 BDI-II CTQ-SF English
Lang et al., 2010 44 0 61 Community/volunteer 29.27 BDI-II CTQ-SF English
Langhinrichsen-Rohling et al., 2011 1533 52 37 Community/volunteer 15.8 CES-D CTQ-SF English
Leenarts et al., 2013 154 0 51 Any clinic referred 16.0 TSCC CTQ-SF Dutch
Leeson & Nixon, 2011 50 46 94 Any clinic referred 11.18 CDI CTQ-SF English
Lehavot et al., 2014 699 0 85 Community/volunteer 49.74 PHQ-8 CTQ-SF English
Levine & Fritz, 2016 51 0 57 Other 37 BDI-II CTQ-SF English
Lewis et al., 2006 102 0 NS Other 27.17 CES-D NS English
Liu et al., 2013 66 23 62 Community/volunteer 19.86 BDI-II CTQ-SF English
Locke et al., 2007 904 0 0 Community/volunteer 17 Measure of dysphoria CTQ-SF NS
Lopez et al., 2011 813 0 40 Other 15.09 CES-D CTQ-SF English
Lowe et al., 2016 3192 30 NS Other 39.98 BDI-II CTQ-SF English
Lu et al., 2016 80 43 NS Any clinic referred 22.59 SCID CTQ-SF NS
MacDonald et al., 2015 200 54 NS Any clinic referred 35.00 BDI-FS, PHQ-9 CTQ-SF English
Malykhin et al., 2010 73 23 85 Community/volunteer 33.91 SCID-I NS English
Marquee-Flentje, 2017 300 0 56 Community/volunteer 26.3 SCL-90-R CTQ-SF English
Martinez-Torteya et al., 2014 153 0 56 Community/volunteer 29.06 PPDS CTQ-SF English
Martsolf, 2004 258 34 NS Community/volunteer 32.4 CES-D CTQ-SF Creole
Massing-Schaffer et al., 2015 185 25 56 Community/volunteer 19.65 BDI-II CTQ-SF English
Mazzeo et al., 2008 (African American) 192 0 0 Community/volunteer 20.15 CES-D CTQ-SF English
Mazzeo et al., 2008 (European American) 412 0 100 Community/volunteer 19.59 CES-D CTQ-SF English
McGinn et al., 2005 55 11 27 Any clinic referred 41.9 BDI CTQ-SF English
McGinnis et al., 2015 198 0 NS Community/volunteer NS PPDS CTQ-SF English
Mehta et al., 2014 62 0 85 Any clinic referred 33.38 BDI, SCID-I, HAMD, EPDS NS English
Michopoulos et al., 2015 1110 20 3 Other 39.6 BDI-II NS English
Mikaeili et al, 2013 893 100 NS Population-based/epidemiological 13.24 SCL-90-R CTQ-SF NS
Miller et al., 2017 682 NS 62 Community/volunteer 11.83 CDI CTQ-SF English
Minnich et al., 2017 1344 36 90 Community/volunteer 18.97 BDI-II CTQ-SF English
Mitchell & Mazzeo, 2005 168 100 54 Community/volunteer 19.7 CES-D CTQ-SF English
Morelen et al., 2016 192 0 59 Any clinic referred 28.88 PPDS CTQ-SF English
Mullins et al., 2016 512 34 NS Any clinic referred 38.61 BDI, SCAN, Past History Schedule CTQ-SF English
Murphy et al., 2012 90 37 NS Any clinic referred 39.35 SCID NS English
Muzik et al., 2017 183 0 59 Other 29.15 PPDS CTQ-SF English
Negele et al., 2015 349 32 NS Any clinic referred 40.40 BDI-II CTQ-SF German
Ng et al., 2011 160 32 0 Any clinic referred 41.9 BDI-II, prior diagnosis CTQ-SF Chinese
Norton, 2017 188 11 67 Any clinic referred NS PROMIS Depression CTQ-53 English
O’Mahen et al., 2015 140 0 49 Other 26.71 BDI-II CTQ-SF English
Opel et al., 2014 170 38 NS Any clinic referred 37.4 BDI, SCID-I CTQ-SF German
Opel et al., 2016 76 50 NS Any clinic referred 36.89 SCID-I CTQ-SF German
Peeters et al., 2002 25 40 NS Any clinic referred 41.5 MADRS CTQ-SF Dutch
Peh et al., 2017 108 41 NS Any clinic referred 17.0 PHQ-8 CTQ-SF NS
Peng et al., 2014 109 53 NS Any clinic referred 28.37 HAMD CTQ-SF Chinese
Philippe et al., 2011 118 30 NS Any clinic referred 32.82 BDI CTQ-SF NS
Pieritz et al., 2015 62 0 NS Community/volunteer 34.4 PHQ-9 CTQ-SF German
Powers et al., 2009 378 46 4 Other 43.1 BDI-II CTQ-SF English
Raab et al., 2012 (female) 56 0 25 Any clinic referred 49.41 MDI CTQ-SF English
Raab et al., 2012 (male) 61 100 15 Any clinic referred 46.68 MDI CTQ-SF English
Raes & Hermans, 2008 101 18 100 Community/volunteer 19.64 BDI CTQ-53 Dutch
Rezaei et al., 2016 439 0 NS Other 22.47 BDI-II CTQ-SF Persian
Rieder & Elbert, 2013 188 47 NS Community/volunteer 21.3 HSCL-25 CTQ-SF Kinyarwanda
Riggs & Kaminski, 2010 285 23 69 Community/volunteer 21.9 HSCL-25 CTQ-SF English
Rikhye et al., 2008 141 35 NS Community/volunteer 31.27 IDS-SR CTQ-SF English
Ritschel et al., 2015 1050 24 42 Community/volunteer 20.66 DASS CTQ-SF English
Salah, 2015 22 5 NS Community/volunteer 19.41 BDI CTQ-SF Dutch
Salwen & Hymowitz, 2015 382 43 50 Community/volunteer 19.26 QIDS CTQ-SF English
Savitz et al., 2008 114 43 100 Any clinic referred 48.8 SCID NS English
Schulz et al., 2014 2265 47 NS Population-based/epidemiological 46.32 BDI-II CTQ-SF German
Schumm et al., 2005 176 0 38 Other 22.10 CES-D CTQ-SF English
Sexton et al., 2015 214 0 61 Community/volunteer 28.2 PDSS CTQ-SF English
Shahar et al., 2015 219 50 NS Community/volunteer 38.7 DASS CTQ-SF NS
Shapero et al., 2013 216 42 53 Community/volunteer 1 CDI CTQ-SF English
Shea et al., 2007 66 0 NS Any clinic referred 30.50 MADRS, MINI, EPDS CTQ-SF English
Shenk et al., 2017 220 0 81 Other 21.26 BDI-II CTQ-SF English
Shi, 2013 497 35 NS Any clinic referred 27.7 TSI CTQ-SF English
Song et al., 2016 305 43 NS Any clinic referred 37.0 BDI CTQ-SF Korean
Specht et al., 2009 117 0 71 Other 33.9 BDI-II CTQ-SF English
Spertus et al., 2003 205 0 80 Other 44.5 SCL-90-R CTQ-SF English
Spinhoven et al., 2014 2308 34 NS Any clinic referred 46.0 IDS CTQ-SF Dutch
Stacks et al., 2014 83 0 73 Community/volunteer 30.04 PPDS CTQ-SF English
Stange et al., 2014 (male) 118 100 NS Community/volunteer 12.32 CDI CTQ-SF English
Stange et al., 2014 (female) 138 0 NS Community/volunteer 12.32 CDI CTQ-SF English
Steffey, 2012 207 27 87 Community/volunteer 21.86 CES-D CTQ-SF English
Stewart et al., 2015 163 23 76 Any clinic referred 15.60 CES-D CTQ-SF English
Suliman et al., 2009 922 41 31 Population-based/epidemiological 15.73 BDI CTQ-SF NS
Sullivan et al., 2012 143 0 9 Community/volunteer 38.09 CES-D CTQ-SF English
Suzuki et al., 2014 79 35 80 Any clinic referred 48.26 Prior diagnosis, QIDS CTQ-SF English
Tanaka et al., 2011 117 45 27 Other 18.1 CES-D CTQ-SF English
Tatham et al., 2016 61 NS 100 Any clinic referred 35.61 SCID-I, HDRS CTQ-SF English
Tlapek et al., 2017 237 0 25 Other 14.9 CDI CTQ-SF English
Tollenaar et al., 2017 2567 34 NS Any clinic referred 42.18 CIDI CTQ-SF English
Tozzi et al., 2016 83 35 NS Any clinic referred 38.80 SCID-I CTQ-SF English
Treadway et al., 2009 38 47 NS Any clinic referred 32.75 SCID, HDRS CTQ-SF English
Ugwu et al., 2015 92 40 NS Any clinic referred 38.24 SCID-I CTQ-SF English
Van der Kloet et al., 2012 266 50 NS Any clinic referred 44.2 BDI-II CTQ-SF Dutch
Van Vugt et al., 2014 89 0 NS Any clinic referred 19.27 TSCC CTQ-SF NS
Virkler, 2006 75 0 96 Community/volunteer 62.75 BDI-II CTQ-SF English
Voth Schrag et al., 2017 105 0 41 Any clinic referred 14.9 CDI CTQ-SF English
Walsh et al., 2016 133 0 NS Other 17.80 SCL-90-R CTQ-SF English
Wanklyn et al., 2012 110 61 31 Other 16.78 CES-D CTQ-SF English
Watson et al., 2007 10b 37 NS Any clinic referred 37.72 prior diagnosis CTQ-SF English
Wessel et al., 2001** 117 46 NS Any clinic referred 36.28 SCID, SDS CTQ-53 Dutch
Wilbertz et al., 2010 32 50 NS Any clinic referred 43.72 BDI CTQ-SF German
Wingenfeld et al., 2017 143 0 NS Any clinic referred 34.77 SCID-I CTQ-SF German
Wingenfeld et al., 2013 36 18 NS Any clinic referred 35.19 SCID-I CTQ-SF German
Wingo et al., 2010 792 32 NS Other 36 BDI CTQ-SF English
Woods et al., 2010 157 0 46 Community/volunteer 33.7 TSI CTQ-SF English
Wu et al., 2018 358 37 NS Community/volunteer 19.18 TDS CTQ-SF Chinese
Wuest et al., 2010 309 0 76 Community/volunteer 39.4 CES-D CTQ-SF English
Yang et al., 2017 168 27 NS Any clinic referred 30.64 SCID-I, HAMD CTQ-SF Chinese
Yildiz Inanici et al., 2017 144 0 NS Other 29.37 BDI CTQ-SF Turkish
Zalewski et al., 2013 95 0 77 Any clinic referred 44 QIDS CTQ-SF English
Open in a new tab

Note. NS = not specified. BDI = Beck Depression Inventory. BDI-FS = Beck Depression Inventory-Fast Screen. BDI-II = Beck Depression Inventory, 2nd edition. BSI = Brief Symptom Inventory. CDI = Children’s Depression Inventory. CES-D = The Center for Epidemiologic Studies Depression Scale. DASS = Depression Anxiety Stress Scales. DIS = Diagnostic Interview Schedule. EPDS = Edinburgh Postnatal Depression Scale. HAMD = Hamilton Depression Rating Scale (also known as HRSD = Hamilton Depression Rating Scale and HRSD = Hamilton Rating Scale for Depression). HSCL-25 = Hopkins Symptom Checklist-25. IDS-SR = The Inventory of Depression Symptomatology, Self-Report. MADRS = Montgomery-Asberg Depression Scale. MDI = Major Depression Index. MINI = The Mini International Neuropsychiatric Interview. PDSQ = Psychiatric Diagnostic Screening Questionnaire. PHQ = Patient Health Questionnaire. PDSS = Postpartum Depression Screening Scale (also known as PPDS = Postpartum Depression Screening Scale). PROMIS = Patient-Reported Outcomes Measurement Information System. QIDS = Quick Inventory of Depression Symptomatology. RADS-2 = Reynolds Adolescent Depression Scale, Second Edition. RCADS = Revised Children’s Anxiety and Depression Scale. SCAN = Schedules for Clinical Assessment in Neuropsychiatry. SCID = Structured Clinical Interview. SCID-I = Structured Clinical Interview for Axis I Disorders. SCID-I/P = Structured Clinical Interview for Axis I Disorders, Patient Edition. SCL-90-R = Symptom Checklist-90-Revised. SDS = Zung Self-Rating Depression Scale. SMFQ = Short Mood and Feelings Questionnaire. TDS = trait depression subscale of the State-Trait Depression Questionnaire. TSCC = Trauma Symptom Checklist for Children. TSI = Trauma Symptom Inventory.

a

Provided 162 participants for the depression scores analyses, presented here, and a subset (104) for the diagnostic group analysis (42% male, mean age = 42.40).

b

Compared 10 individuals with MDD to 1000 individuals from a population representative sample.

**

Provided 117 participants for the diagnostic group analysis, presented here, and a subset (91) for the depression scores analyses (45% male, mean age = 36.60).

Child Maltreatment and Continuous Depression Scores

The number of studies that examined the relation between severity of child maltreatment and depression scores was 70 for total CTQ scores, and ranged from 48 (physical neglect) to 81 (emotional abuse) for the subtype CTQ scores. Overall, there was a significant association between child maltreatment and depressive symptoms (Figure 2). The effect size estimates varied by type of child maltreatment: estimates were highest for emotional abuse and lowest for sexual abuse. All effect sizes differed significantly from 0, indicating a significant association between all types of child maltreatment and depression scores. Variation of the effect size within each meta-analysis is presented in Table 2. In addition, there was evidence of significant heterogeneity for all outcomes.

Figure 2.

Figure 2.

Open in a new tab

Estimated association (Z) between total childhood trauma questionnaire scores and depressive symptoms. Estimates of zero indicate no association. Positive values indicate a positive association between maltreatment scores and continuous depression scores.

Child Maltreatment and Depression Diagnosis

The number of studies that examined the relation between severity of child maltreatment and diagnosis of MDD as 39 for total CTQ scores and 35 for each of the subtype CTQ scores. As with the correlational analyses, there was a significant association between total CTQ scores and a diagnosis of depression (Figure 3). The random-effects meta-analysis indicated that individuals with depression reported higher child maltreatment scores than did individuals without depression, an effect that differed significantly from zero. The effect size estimates varied by type of child maltreatment: they were highest for emotional neglect and lowest for sexual abuse. All effect sizes obtained from meta-analyses differed significantly from zero, indicating a significant association between all types of child maltreatment and depression scores. Variation of the effect size within each meta-analysis is presented in Table 3. In addition, there was evidence of significant heterogeneity for all outcomes.

Figure 3.

Figure 3.

Open in a new tab

Estimated standardized mean difference (Hedge’s g) in childhood trauma questionnaire total scores between individuals with and without a diagnosis of depression. Estimates of zero indicate no differences, whereas an effect size of one indicates a full standard deviation difference in scores. Positive values indicate higher scores among those with a diagnosis of depression.

Moderators

We examined both methodological and demographic study-level variables that may explain variation in effect sizes within studies for each outcome (see Method for moderator variables of interest). We tested each coded moderator separately using simple regressions, weighted by the sample size for each study. Statistically significant moderators are presented by outcome in Tables 2 and 3. For total CTQ score and depressive symptoms, community samples were associated with larger effect size relative to other participant sources (t(69) = 3.18, p = .002). When the studies were divided based on whether the samples were drawn from the community vs. all others (e.g., clinic, population-based, etc.), we observed that the 48 samples not drawn from the community had a statistically significant (Z = 16.08, p < .001), but somewhat smaller estimate of the effect size (Z = .32 [95% CI, .28−.36]) than did the 22 studies of community participants (Z = .43 [95% CI, .37−.49]), whose overall effect statistically differed from zero (Z = 14.06, p < .001). In addition, studies that used the CES-D, relative to other measures (e.g., BDI, etc.), on average had larger effect sizes (t(69) = 2.34, p = .022). Again, when the studies were divided based on the depression assessment measure, we found that studies that used the CES-D (n = 11) had a larger effect size estimate (Z = .43 [95% CI, .36−.51]) than did studies that did not use the CES-D (n = 59) (Z = .34 [95% CI, .30−.37]), although both sets of studies had effects that differed significantly from zero).

For emotional abuse, whether the mean age of the sample fell into childhood or adulthood (i.e., split based on the mean age of 18 years) emerged as a significant moderator (t(80) = 2.34, p = .022). Analyses conducted within the 20 studies with child/adolescent samples yielded a larger association (Z = .45 [95% CI, .35−.54]) than did the 59 studies that included adults (Z = .36 [95% CI, .32−.39]), although in both cases the estimates significantly differed from zero (Z = 9.28, p < .001 and Z = 20.86, p < .001, respectively) and remained significantly heterogeneous. For this outcome, sample source was also significantly associated with effect size, such that population-based samples had smaller effect sizes than did other sample sources (t(80) = −3.12, p = .003). When the two studies that were population-based (i.e., Mikaeili, Barahmand, & Abdi, 2013; Schulz, Schmidt, et al., 2014) were excluded, the overall effect was similar to the full analyses (Z = .38 [95% CI, .35−.41]) and the effect statistically differed from zero (Z = 24.72, p < .001). In addition, year of publication was significantly associated with effect size (Coef. = 0.01, SE = 0.004; t(80) = 2.35, p = .021): on average, more recently published papers had larger effects. For both emotional neglect and physical neglect, population-based samples had smaller effect sizes than did other sample sources (emotional neglect: Coef. = −0.21, SE = 0.07; t(57) = −2.87, p = .006; physical neglect: Coef. = −0.21, SE = 0.10; t(47) = −2.04, p = .047). When population-based samples were excluded, the overall effect was just slightly larger relative to the full analyses (emotional neglect: Z = .31 [95% CI, .27−.34], physical neglect: Z = .24 [95% CI, .20−.27]); the effects of non-population-based samples on depression differed statistically from zero (emotional neglect: Z = 17.24, p<.001, physical neglect: Z = 14.34, p<.001).

Finally, for the depression group analyses, a significantly larger effect size was found in studies that used the full 53-item version of the CTQ to assess the association between total CTQ score and depression group (t(38) = −2.62, p = .013). Analyses were repeated in the 11 studies that used the original version and the 28 studies that used the short form; in both sets, the effect size estimates differed significantly from zero (CTQ-53: g = 1.64 [95% CI, 1.14−2.14], Z = 6.45, p < .001 and CTQ-SF: g = 0.93 [95% CI, 0.82−1.05], Z = 15.59, p < .001). In addition, for CTQ total scores, the language in which the measure was administered moderated the observed effect size: studies conducted in English had smaller effect sizes than did non-English studies (Coef. = 0.50, SE = 0.22, t(38) = 2.23, p = .032); both English and non-English studies had effect size estimates that differed significantly from zero (English: g = 1.40 [95% CI, [1.10, 1.70], Z = 9.09 p < .001; non-English: g = 0.89 [95% CI, [0.80, 0.99], Z = 18.45, p < .001). The same pattern was found for both sexual abuse (English: g = 0.65 [95% CI, [0.48, 0.82], Z = 7.40, p < .001; non-English: g = 0.37 [95% CI, [0.31, 0.42], Z = 12.46, p < .001) and emotional neglect (English: g = 1.17 [95% CI, [0.93, 1.42], Z = 9.29,p < .001; non-English: g = 0.85 [95% CI, [0.73, 0.96], Z = 14.16, p < .001).

Publication Bias

For CTQ total score and depressive symptoms, the Egger’s test revealed statistically significant bias (t(69) = −2.02, p = .047). The negative intercept (Coef. = −0.91, SE = 0.45) indicates that the effects from the smaller studies are less than the effects from the larger studies, indicating that small studies are not upwardly biasing the estimate. A trim and fill procedure identified 0 missing studies. For emotional abuse, there was evidence of publication bias. The Egger’s test was statistically significant (t(80) = 2.41, p = .018), with a positive intercept (Coef. = 1.42, SE = 0.59) indicating that smaller studies may be upwardly biasing the effect. A trim and fill procedure identified 27 missing studies, with a filled meta-analysis estimate of Z = .29 (95% CI, .25−.33).

For the group-based analyses, there was evidence of publication bias from Egger’s test for total CTQ scores (Coef. = 1.31, SE = 0.56, t(34) = 2.32, p = .026), emotional neglect (Coef. = 1.40, SE = 0.45, t(34) = 3.12, p = .004), physical neglect (Coef. = 1.52, SE = 0.51, t(34) = 2.97, p = .005). In all cases, smaller studies may have been upwardly biasing estimates. Trim and fill procedures indicated the following corrected effect size estimates for total CTQ scores: g = .88 (95% CI, 0.74−1.02; 8 missing), emotional neglect; g = .77 (95% CI, 0.65−0.90; 12 missing), and physical neglect: g = .49 (95% CI, 0.35−0.63; 10 missing). In all cases, these revised estimates had effects that differed significantly from zero. No other associations were characterized by statistically significant tests of publication bias.

Leave-one-out Sensitivity Analyses

Given the significant heterogeneity in effects, we conducted sensitivity analyses for all of the outcomes using the leave-one-out approach (i.e., conducting the random-effects model following the removal of each study individually, with replacement). Tables 2 and 3 provide data indicating that no single study unduly influenced the effect size estimates; in all cases in which a study was removed, the effect size estimates remained significantly different from zero.

Discussion

In this paper we report the result of a meta-analysis of 192 unique samples from 190 studies, consisting of 68,830 individuals, conducted to test whether child maltreatment was associated with depression diagnosis and symptom scores in adulthood. This is the largest study examining the association between child maltreatment and depression, increasing our confidence in the strength of the observed effect sizes. Across both methods of assessing depressive symptomatology, we found a significantly increased risk for higher depression symptom scores and depressive disorders (typically meeting criteria for MDD) as a function of greater reported severity of child maltreatment. In addition, in order to examine whether there was specificity in the association between different types of child maltreatment and depression, we conducted analyses across five types of maltreatment, all assessed using the same measure of child maltreatment (i.e., the CTQ). Consistent with expectations, we found that all types of maltreatment were associated with significantly higher depression scores and greater risk for meeting criteria for MDD. Importantly, however, emotional abuse and emotional neglect had the strongest associations with depression; we found weaker associations for sexual and physical abuse and physical neglect. In addition, the magnitude of the effect between emotional abuse and depressive symptoms was larger in samples of children and adolescents than in samples of adults.

The estimated effect size between child maltreatment scores and later depression was large; specifically, individuals with depression had, on average, total child maltreatment scores that were approximately one standard deviation higher than scores of their nondepressed counterparts. Even after applying a trim-and-fill procedure following the identification of possible publication bias favoring smaller studies with larger effects, the estimated effect size was almost one standard deviation difference between groups. These effects are substantially larger than those previously reported, which is bolstered by the large number of unique individuals who contributed data to these analysis and the advances in methods by including a dimensional assessment of child maltreatment. For example, across 9 studies, a composite measure of childhood maltreatment was reported to be moderately associated with a diagnosis of depression, although the confidence interval included zero (SMD=0.43; Infurna et al., 2016). It is possible that this discrepancy is due to differences in the scales used in these two meta-analyses (CTQ vs. CECA); for example, the range of possible scores is substantially greater in the CTQ and, further, many of the studies in Infurna et al.’s study used dichotomized experiences of maltreatment rather than applying a dimensional approach to assessing maltreatment.

Although all forms of child maltreatment examined in the present study were significantly associated with depression, the strength of the association varied by type of maltreatment. Three prior meta-analyses are relevant in interpreting these findings. Mandelli, Petrelli, and Serretti (2015) meta-analyzed studies that examined the association between binary measures of child maltreatment and diagnosed depression. These investigators found that emotional abuse (k=8) and neglect (k=6) were most strongly associated with depression (ORs=2.8), and reported a weaker association for physical abuse (k=10; OR=2.0). Infurna et al. (2016) found that psychological abuse and neglect were the types of maltreatment most strongly associated with depression, and reported weaker, although still statistically significant, associations for sexual abuse. Finally, Norman et al. (2012) examined three forms of child maltreatment in relation to depressive disorders, and found the strongest association with emotional abuse (OR=3.06), followed by neglect, broadly defined (OR=2.11), and the weakest association for physical abuse (OR=1.54). While all three effect estimates differed significantly from zero, the effect estimate for emotional abuse and depression was significantly stronger than that for physical abuse and depression. In the present study, although effect size estimates varied across types of maltreatment, for depression diagnosis we found that emotional abuse differed significantly from physical abuse and sexual abuse, and that emotional neglect differed significantly from physical abuse, sexual abuse, and physical neglect, as represented by non-overlapping CIs. Such findings suggest that, for depression, predictions are less informed by whether maltreatment experiences are characterized by threat versus deprivation (e.g., McLaughlin et al., 2014; Sheridan & McLaughlin, 2014); instead, emotional maltreatment in particular could be depressogenic.

Importantly, more “silent” forms of child maltreatment (i.e., emotional abuse and emotional neglect) are most strongly associated with depression. This finding is consistent with theoretical and empirical accounts of maltreatment and depression. Compared to sexual and physical abuse, emotional neglect has been found to be uniquely associated with anhedonic symptoms of depression (Van Veen et al., 2013). Furthermore, Rose and Abramson’s (1992) developmental extension of the hopelessness theory of depression provides a framework through which to view the potential differential effects of emotional maltreatment. Rose and Abramson hypothesized that emotional abuse leaves individuals particularly vulnerable to developing a negative cognitive style, which in turn increases risk for depression. According to this formulation, children seek to understand the cause of the adverse life events they experience. Initially, these explanations are external, unstable, and specific (e.g., concluding that the cause of the abuse is not due to their stable characteristics of themselves, but instead, to some outside, isolated reason—for example, a parent having a stressful day). However, in the case of recurrent abuse, children may develop a more depressogenic causal attribution for the abuse (i.e., an attribution that is internal, stable, and global). In this context, emotional abuse may be particularly detrimental to children’s cognitive style because the abuser may state the negative causal attribution to the child (e.g., being called names). This formulation is supported by empirical work: emotional maltreatment during childhood has been found to be associated with negative self-referential processing (Steinberg, Gibb, Alloy, & Abramson, 2003), one potential risk pathway for depression. Our findings suggest that emotional neglect plays a similarly harmful role; thus, a depressive cognitive style may stem not only from the communication of negative cognitions, as in the case of emotional abuse, but also from lack of emotional support, as is the case with emotional neglect.

Among the moderators examined in this study, sample source and language used (i.e., English vs. other languages) emerged as particularly salient in relation to the size of effects that were estimated. Specifically, population-based studies demonstrated smaller associations than did other study recruitment sources. Sample source has also been found to be a relevant moderator in other studies; our findings are similar to those documenting a stronger association between child maltreatment and depression in clinical than in population-based samples (Infurna et al., 2016). Type of sample (i.e., community vs. clinical) has been found to be associated with type of maltreatment and risk for depression (Mandelli et al., 2015); across sample types, these investigators found a strong association between neglect and depression; in contrast, in community samples emotional abuse was a stronger predictor of depression. In addition, studies conducted in English had larger effect sizes, on average, than did those conducted in other languages. Language is confounded with geography and cultural factors, and it is difficult to disentangle which of these may be responsible for explaining the differences in effect size based on this moderator. In addition, we found significant evidence of publication bias in several of the meta-analyses here. Our analyses conducted to identify what is more likely to be unbiased effects all continue to demonstrate a significant association between maltreatment and depression, although the effect sizes are lower and are more likely to be an accurate estimate of the magnitude of the associations.

Despite the plausibility that other moderators (e.g., sex, age of participants) are meaningful in understanding the link between child maltreatment and depression, for nearly all outcomes we found no significant evidence that the size of the effect was explained by these factors. For emotional abuse, however, we did find evidence of a larger effect in the relation to child depressive symptoms than those found in adult samples. Such findings may indicate that the association between emotional abuse and depression symptoms weakens over time and as individuals enter adulthood.

We should note five limitations of the present meta-analysis. First, because the studies analyzed for this meta-analysis were cross-sectional, we cannot speak to a direct causal link between emotional maltreatment and depression. In this context, there may be gene-environment correlations for depression and maltreatment, given that parents with depression not only are passing on their genes, but also are more likely to engage in child maltreatment (Widom, DuMont, & Czaja, 2007). Second, the types of child maltreatment assessed in this meta-analysis do not occur independent of one another. Experiences of maltreatment, as well as other forms of stress in early life, are not randomly distributed: children who experience any one type of maltreatment are more likely to have experienced other types (Edwards et al., 2003). We are unable to determine the independent effect of each type using this approach, and we believe it would be useful going forward to use dimensional assessments of maltreatment type to document more thoroughly the overlap between each form of maltreatment. Third, we did not require that studies conduct clinical assessments to make diagnostic determinations of depression. While there are strengths to assessing depression dimensionally (see Ruscio & Ruscio, 2000), the use of clinical instruments may better capture functional impairment in relation to depression. Fourth, the assessment of child maltreatment in these studies was retrospective. While prospective studies also support the link between child maltreatment and depression (Li et al., 2016), there may be selective or biased reporting of adversity, which could affect the observed nature of the association between child maltreatment and depression (Colman et al., 2016; Patten et al., 2015). In fact, recent meta-analyses indicate that prospective and retrospective reports of maltreatment may identify different subgroups of individuals (Baldwin, Reuben, Newbury, & Danese, 2019), which could mean that what we documented here as potential predictive pathways may instead be a better marker of concurrent mood and recollections of past experiences. Finally, the CTQ is not without flaws. We selected this measure given its widescale use, its ability to assess maltreatment using a dimensional approach, and its assessment of different subtypes of maltreatment. However, the CTQ does not provide details about the timing of events, which are likely to be important in understanding the association between stress and depression (Teicher, 2008). It also has psychometric limitations. In particular, researchers have noted low reliability of the physical neglect subscale (Gil et al., 2009; Paivio & Cramer, 2004), that has been attributed to greater variability in the types of items included on this subscale (Bernstein et al., 1994).

In closing, the present findings underscore the association between experiences of child maltreatment and depression in adulthood. The goal of the present study was to characterize the associations between depression in adulthood and child maltreatment generally, as well as specific forms of child maltreatment. Assuming that there is a causal link between child maltreatment and depression, next steps in this line of research include probing the potential mechanisms by which these early adverse experiences may lead to a diagnosis of depression and to increased levels of depressive symptomatology in adulthood. Identifying these mechanisms will be important in understanding why treatment response has been found to be moderated by childhood maltreatment status, with individuals who endorsed child maltreatment being less likely to respond to treatment (Nanni et al., 2012). Collectively, these results highlight the importance of reducing exposure to child maltreatment as a clear policy goal. Interventions and preventions that have been shown to reduce child maltreatment are important, and include the Nurse Family Partnership (Donelan-McCall, Eckenrode, & Olds, 2009) and the Triple P (Positive Parenting Program) (Prinz, Sanders, Shapiro, Whitaker, & Lutzker, 2009). While there is likely to be immediate benefit for the children and the parents who participate in these programs, it is notable that the effects may also have long-term positive mental health outcomes (Liu, 2017). Finally, researchers must consider emotional maltreatment (i.e., emotional abuse and emotional neglect) as influencing the etiology of depression; indeed, including these more silent forms of maltreatment in relevant studies should yield important insights concerning the causes of depression and treatment targets for individuals who are experiencing this debilitating disorder.

Acknowledgements

This work was supported by the National Institutes of Health (R37-MH101495 to IHG and F32-MH107129 to KLH), the Stanford Precision Health and Integrated Diagnostics (PHIND) Center to IHG, the Brain & Behavior Research Foundation (NARSAD Young Investigator [23819 to KLH and 22337 to JL]), the Social Sciences and Humanities Research Council (430-2017-00408 to JL), the Canadian Institute of Health Research (389703 to JL), the Klingenstein Third Generation Foundation Fellowship (to KLH), and the Jacobs Foundation Early Career Research Award (to KLH).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Aguilera M, Arias B, Wichers M, Barrantes-Vidal N, Moya J, Villa H, … Fañanás L (2009). Early adversity and 5-HTT/BDNF genes: new evidence of gene-environment interactions on depressive symptoms in a general population. Psychological Medicine, 39(9), 1425–1432. 10.1017/S0033291709005248 [DOI] [PubMed] [Google Scholar]
  2. Akbaba Turkoglu S, Essizoglu A, Kosger F, & Aksaray G (2015). Relationship between dysfunctional attitudes and childhood traumas in women with depression. International Journal of Social Psychiatry, 61(8), 796–801. 10.1177/0020764015585328 [DOI] [PubMed] [Google Scholar]
  3. Aversa LH, Lemmer J, Nunnink S, McLay RN, & Baker DG (2014). Impact of childhood maltreatment on physical health-related quality of life in U.S. active duty military personnel and combat veterans. Child Abuse and Neglect, 38(8), 1382–1388. https://doi.Org/10.1016/j.chiabu.2014.03.004 [DOI] [PubMed] [Google Scholar]
  4. Bailer J, Witthöft M, Wagner H, Mier D, Diener C, & Rist F (2014). Childhood maltreatment is associated with depression but not with hypochondriasis in later life. Journal of Psychosomatic Research, 77(2), 104–108. 10.1016/j.jpsychores.2014.06.004 [DOI] [PubMed] [Google Scholar]
  5. Baldwin JR, Reuben A, Newbury JB, & Danese A (2019). Agreement between Prospective and Retrospective Measures of Childhood Maltreatment: A Systematic Review and Meta-analysis. JAMA Psychiatry. 10.1001/jamapsychiatry.2019.0097 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Banducci AN, Hoffman E, Lejuez CW, & Koenen KC (2014). The relationship between child abuse and negative outcomes among substance users: Psychopathology, health, and comorbidities. Addictive Behaviors, 39(10), 1522–1527. 10.1016/j.addbeh.2014.05.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bernet CZ, & Stein MB (1999). Relationship of childhood maltreatment to the onset and course of major depression in adulthood. Depression and Anxiety, 9(4), 169–174. [PubMed] [Google Scholar]
  8. Bernstein DP, & Fink L (1998). childhood trauma questionnaire: A retrospective self-report: Manual. Harcourt Brace & Company. [Google Scholar]
  9. Bernstein DP, Fink L, Handelsman L, Foote J, Lovejoy M, Wenzel K, … Ruggiero J (1994). Initial reliability and validity of a new retrospective measure of child abuse and neglect. American Journal of Psychiatry. 10.1176/ajp.151.8.1132 [DOI] [PubMed] [Google Scholar]
  10. Bernstein DP, Stein JA, Newcomb MD, Walker E, Pogge D, Ahluvalia T, … Zule W (2003). Development and validation of a brief screening version of the Childhood Trauma Questionnaire. Child Abuse & Neglect, 27(2), 169–190. 10.1016/S0145-2134(02)00541-0 [DOI] [PubMed] [Google Scholar]
  11. Bifulco A, Brown GW, & Harris TO (1994). Childhood experience of care and abuse (CECA): A retrospective interview measure. Journal of Child Psychology and Psychiatry, 35(8), 1419–1435. 10.1111/j.1469-7610.1994.tb01284.x [DOI] [PubMed] [Google Scholar]
  12. Colman I, Kingsbury M, Garad Y, Zeng Y, Naicker K, Patten S, … Thompson AH (2016). Consistency in adult reporting of adverse childhood experiences. Psychological Medicine, 46(3), 543–549. [DOI] [PubMed] [Google Scholar]
  13. Donelan-McCall N, Eckenrode J, & Olds DL (2009). Home visiting for the prevention of child maltreatment: Lessons learned during the past 20 years. Pediatric clinics of North America, 56(2), 389–403. 10.1016/j.pcl.2009.01.002 [DOI] [PubMed] [Google Scholar]
  14. Edwards VJ, Holden GW, Felitti VJ, & Anda RF (2003). Relationship between multiple forms of childhood maltreatment and adult mental health in community respondents: Results from the adverse childhood experiences study. The American Journal of Psychiatry, 160(8), 1453–1460. 10.1176/appi.ajp.160.8.1453 [DOI] [PubMed] [Google Scholar]
  15. Egger M, Davey Smith G, Schneider M, Minder C, Mulrow C, Egger M, … Olkin I (1997). Bias in meta-analysis detected by a simple, graphical test. British Medical Journal (Clinical Research Ed.), 315(7109), 629–634. 10.1136/bmj.315.7109.629 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fuller-Thomson E, Battiston M, Gadalla TM, & Brennenstuhl S (2014). Bouncing back: Remission from depression in a 12-year panel study of a representative Canadian community sample. Social Psychiatry and Psychiatric Epidemiology, 49(6), 903–910. 10.1007/s00127-013-0814-8 [DOI] [PubMed] [Google Scholar]
  17. Gil A, Gama CS, de Jesus DR, Lobato MI, Zimmer M, & Belmonte-de-Abreu P (2009). The association of child abuse and neglect with adult disability in schizophrenia and the prominent role of physical neglect. Child Abuse and Neglect. 10.1016/j.chiabu.2009.02.006 [DOI] [PubMed] [Google Scholar]
  18. Güleç MY, Altinta M, Inan9 L, Bezgin ÇH, Koca EK, & Güleç H (2013). Effects of childhood trauma on somatization in major depressive disorder: The role of alexithymia. Journal of Affective Disorders, 146(1), 137–141. 10.1016/j.jad.2012.06.033 [DOI] [PubMed] [Google Scholar]
  19. Hedges LV, & Olkin I (1983). Regression models in research synthesis. The American Statistician, 37(2), 137–140. [Google Scholar]
  20. Hentze C, Walter H, Schramm E, Drost S, Schoepf D, Fangmeier T, … Schnell K (2016). Functional Correlates of childhood maltreatment and symptom severity during affective theory of mind tasks in chronic depression. Psychiatry Research - Neuroimaging, 250, 1–11. 10.1016/j.pscychresns.2016.02.004 [DOI] [PubMed] [Google Scholar]
  21. Humphreys KL, Eng T, & Lee SS (2013). Stimulant medication and substance use outcomes: a meta-analysis. JAMA Psychiatry (Chicago, Ill.), 70(7), 740–749. 10.1001/jamapsychiatry.2013.1273 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Humphreys KL, & Zeanah CH (2015). Deviations from the expectable environment in early childhood and emerging psychopathology. Neuropsychopharmacology, 40(1), 154–170. 10.1038/npp.2014.165 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Infurna MR, Reichl C, Parzer P, Schimmenti A, Bifulco A, & Kaess M (2016). Associations between depression and specific childhood experiences of abuse and neglect: A meta-analysis. Journal of Affective Disorders, 190, 47–55. 10.1016/j.jad.2015.09.006 [DOI] [PubMed] [Google Scholar]
  24. Jonas W, Mileva-Seitz V, Girard AW, Bisceglia R, Kennedy JL, Sokolowski M, … Steiner M (2013). Genetic variation in oxytocin rs2740210 and early adversity associated with postpartum depression and breastfeeding duration. Genes, Brain and Behavior, 12(7), 681–694. 10.1111/gbb.12069 [DOI] [PubMed] [Google Scholar]
  25. Kessler RC, McLaughlin KA, Green JG, Gruber MJ, Sampson NA, Zaslavsky AM, … Williams DR (2010). Childhood adversities and adult psychopathology in the WHO World Mental Health Surveys. The British Journal of Psychiatry, 197(5), 378–385. 10.1192/bjp.bp.110.080499 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. King LS, Humphreys KL, & Gotlib IH (2019). The neglect-enrichment continuum: Characterizing variation in early caregiving environments. Developmental Review, 51, 109–122. 10.1016/j.dr.2019.01.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Klein DN, & Kotov R (2016). Course of depression in a 10-year prospective study: Evidence for qualitatively distinct subgroups. Journal of Abnormal Psychology, 125(3), 337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kounou KB, Bui E, Dassa KS, Hinton D, Fischer L, Djassoa G, … Schmitt L (2013). Childhood trauma, personality disorders symptoms and current major depressive disorder in Togo. Social Psychiatry and Psychiatric Epidemiology, 48(7), 1095–1103. 10.1007/s00127-012-0634-2 [DOI] [PubMed] [Google Scholar]
  29. LeMoult J, Humphreys KL, Tracy A, Hoffmeister J-A, Ip E, & Gotlib IH (2019). Meta-analysis: Exposure to early life stress and risk for depression in childhood and adolescence. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Li M, D’Arcy C, & Meng X (2016). Maltreatment in childhood substantially increases the risk of adult depression and anxiety in prospective cohort studies: systematic review, meta-analysis, and proportional attributable fractions. Psychological Medicine, 46(04), 717–730. 10.1017/S0033291715002743 [DOI] [PubMed] [Google Scholar]
  31. Lipsey MW, & Wilson DB (2001). Practical meta-analysis. Applied Social Research Methods Series, 49, 264 10.1016/j.autneu.2007.06.087 [DOI] [Google Scholar]
  32. Liu RT (2017). Childhood adversities and depression in adulthood: Current findings and future directions. Clinical Psychology: Science and Practice, 24(2), 140–153. 10.1111/cpsp.12190 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lu S, Gao W, Huang M, Li L, & Xu Y (2016). In search of the HPA axis activity in unipolar depression patients with childhood trauma: Combined cortisol awakening response and dexamethasone suppression test. Journal of Psychiatric Research, 78, 24–30. 10.1016/j.jpsychires.2016.03.009 [DOI] [PubMed] [Google Scholar]
  34. MacDonald K, Thomas ML, MacDonald TM, & Sciolla AF (2014). A perfect childhood? Clinical correlates of minimization and denial on the Childhood Trauma Questionnaire. Journal of Interpersonal Violence, 30(6), 0886260514539761 10.1177/0886260514539761 [DOI] [PubMed] [Google Scholar]
  35. Mandelli L, Petrelli C, & Serretti A (2015). The role of specific early trauma in adult depression: A meta-analysis of published literature. Childhood trauma and adult depression. European Psychiatry, 30(6), 665–680. 10.1016/j.eurpsy.2015.04.007 [DOI] [PubMed] [Google Scholar]
  36. McGinnis E, Bocknek E, Beeghly M, Rosenblum KL, & Muzik M (2015). Does Child Sex Moderate Vulnerability to Postpartum Risk among Infants of Mothers at Risk for Psychopathology? Infancy, 20(1), 42–69. [Google Scholar]
  37. McLaughlin KA, & Sheridan MA (2016). Beyond cumulative risk: A dimensional approach to childhood adversity. Current Directions in Psychological Science, 25(4), 239–245. 10.1177/0963721416655883 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. McLaughlin KA, Sheridan MA, & Lambert HK (2014). Childhood adversity and neural development: Deprivation and threat as distinct dimensions of early experience. Neuroscience and Biobehavioral Reviews, 47(11), 578–591. 10.1016/j.neubiorev.2014.10.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Mikaeili N, Barahmand U, & Abdi R (2013). The prevalence of different kinds of child abuse and the characteristics that differentiate abused from nonabused male adolescents. Journal of Interpersonal Violence, 28(5), 975–996. 10.1177/0886260512459377 [DOI] [PubMed] [Google Scholar]
  40. Mullen PE, Martin JL, Anderson JC, Romans SE, & Herbison GP (1996). The long-term impact of the physical, emotional, and sexual abuse of children: A community study. Child Abuse & Neglect, 20(1), 7–21. [DOI] [PubMed] [Google Scholar]
  41. Muscatell KA, Humphreys KL, & Brosso SN (2018). Socioeconomic Status and Inflammation: A Meta-Analysis. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Nanni V, Uher R, & Danese A (2012). Childhood maltreatment predicts unfavorable course of illness and treatment outcome in depression: A meta-analysis. American Journal of Psychiatry, 169(2), 141–151. 10.1176/appi.ajp.2011.11020335 [DOI] [PubMed] [Google Scholar]
  43. Negele A, Kaufhold J, Kallenbach L, & Leuzinger-Bohleber M (2015). Childhood trauma and its relation to chronic depression in adulthood. Depression Research and Treatment, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Norman RE, Byambaa M, De R, Butchart A, Scott J, & Vos T (2012). The long-term health consequents of child physical abuse, emotional abuse, and neglect: A systematic review and meta-analysis. PLoS Med, 9(11), e1001349 10.1371/journal.pmed.1001349 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Paivio SC, & Cramer KM (2004). Factor structure and reliability of the Childhood Trauma Questionnaire in a Canadian undergraduate student sample. Child Abuse and Neglect. 10.1016/j.chiabu.2004.01.011 [DOI] [PubMed] [Google Scholar]
  46. Patten SB, Wilkes TCR, Williams JVA, Lavorato DH, El-Guebaly N, Schopflocher D, … Bulloch AGM (2015). Retrospective and prospectively assessed childhood adversity in association with major depression, alcohol consumption and painful conditions. Epidemiology and Psychiatric Sciences, 24(2), 158–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Peng H, Long Y, Li J, Guo Y, Wu H, Yang Y, … Ning Y (2014). Hypothalamic-pituitary-adrenal axis functioning and dysfunctional attitude in depressed patients with and without childhood neglect. BMC Psychiatry, 14(1), 45 10.1186/1471-244X-14-45 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Petersen AC, Joseph J, & Feit M (2014). New directions in child abuse and neglect research. National Academies Press. [PubMed] [Google Scholar]
  49. Pieritz K, Rief W, & Euteneuer F (2015). Childhood adversities and laboratory pain perception. Neuropsychiatric Disease and Treatment, 11, 2109–2116. 10.2147/NDT.S87703 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Prinz RJ, Sanders MR, Shapiro CJ, Whitaker DJ, & Lutzker JR (2009). Population-based prevention of child maltreatment: The US Triple P system population trial. Prevention Science, 10(1), 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Rhebergen D, Lamers F, Spijker J, De Graaf R, Beekman ATF, & Penninx B (2012). Course trajectories of unipolar depressive disorders identified by latent class growth analysis. Psychological Medicine, 42(7), 1383–1396. [DOI] [PubMed] [Google Scholar]
  52. Rieder H, & Elbert T (2013). The relationship between organized violence, family violence and mental health: Findings from a community-based survey in Muhanga, Southern Rwanda. European Journal of Psychotraumatology, 4 10.3402/ejpt.v4i0.21329 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Rose DT, & Abramson LY (1992). Developmental predictors of depressive cognitive styles. Developmental Perspectives on Depression, 323–349. [Google Scholar]
  54. Ruscio J, & Ruscio AM (2000). Informing the continuity controversy: A taxometric analysis of depression. Journal of Abnormal Psychology, 109(3), 473–487. https://doi.Org/10.1037/0021-843X.109.3.473 [PubMed] [Google Scholar]
  55. Savitz JB, van der Merwe L, Newman TK, Solms M, Stein DJ, & Ramesar RS (2008). The relationship between childhood abuse and dissociation. Is it influenced by catechol-O-methyltransferase (COMT) activity? The International Journal of Neuropsychopharmacology, 11, 149–161. 10.1017/S1461145707007900 [DOI] [PubMed] [Google Scholar]
  56. Schulz A, Becker M, Van Der Auwera S, Barnow S, Appel K, Mahler J, … Grabe HJ (2014). The impact of childhood trauma on depression: Does resilience matter? Population-based results from the Study of Health in Pomerania. Journal of Psychosomatic Research, 77, 97–103. 10.1016/j.jpsychores.2014.06.008 [DOI] [PubMed] [Google Scholar]
  57. Schulz A, Schmidt CO, Appel K, Mahler J, Spitzer C, Wingenfeld K, … Grabe HJ (2014). Psychometric functioning, socio-demogiapic variability of childhood maltreatment in the general population and its effects of depression. International Journal of Methods in Psychiatric Research, 23(3), 387–400. 10.1002/mpr.1447 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Sexton MB, Hamilton L, McGinnis EW, Rosenblum KL, & Muzik M (2015). The roles of resilience and childhood trauma history: Main and moderating effects on postpartum maternal mental health and functioning. Journal of Affective Disorders, 174, 562–568. 10.1016/j.jad.2014.12.036 [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Sheridan MA, & McLaughlin KA (2014). Dimensions of early experience and neural development: deprivation and threat. Trends in Cognitive Sciences, 18(11), 580–585. 10.1016/j.tics.2014.09.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Spinhoven P, Penninx BBW, Hickendorff M, van Hemert AM, Bernstein DP, & Elzinga BM (2014). Childhood Trauma Questionnaire: Factor structure, measurement invariance, and validity across emotional disorders. Psychological Assessment, 26(3), 717 10.1037/pas0000002 [DOI] [PubMed] [Google Scholar]
  61. Steinberg JA, Gibb BE, Alloy LB, & Abramson LY (2003). Childhood emotional maltreatment, cognitive vulnerability to depression, and self-referent information processing in adulthood: Reciprocal relations. Journal of Cognitive Psychotherapy, 17(4), 347–358. [Google Scholar]
  62. Teicher MH (2008). Stress, sensitive periods and maturational events in adolescent depression. Trends in Neurosciences. 10.1016/j.tins.2008.01.004 [DOI] [PubMed] [Google Scholar]
  63. Van Veen T, Wardenaar KJ, Carlier IVE, Spinhoven P, Penninx b., & Zitman FG (2013). Are childhood and adult life adversities differentially associated with specific symptom dimensions of depression and anxiety? Testing the tripartite model. Journal of Affective Disorders, 146(2), 238–245. [DOI] [PubMed] [Google Scholar]
  64. Virkler PM (2006). The relationship between childhood sexual abuse and measures of depression, anxiety and revictimization in females aged 55 to 85., (66(11)), 3942A. [Google Scholar]
  65. Watson S, Owen BM, Gallagher P, Hearn AJ, Young AH, & Ferrier IN (2007). Family history, early adversity and the hypothalamic-pituitary-adrenal (HPA) axis: Mediation of the vulnerability to mood disorders. Neuropsychiatric Disease and Treatment, 3(5), 647–653. [PMC free article] [PubMed] [Google Scholar]
  66. Widom CS, DuMont K, & Czaja SJ (2007). A prospective investigation of major depressive disorder and comorbidity in abused and neglected children grown up. Archives of General Psychiatry, 64(1), 49–56. 10.1016/S0084-3970(08)70770-9 [DOI] [PubMed] [Google Scholar]
  67. Wingo A. p., Wrenn G, Pelletier T, Gutman AR, Bradley B, & Ressler KJ (2010). Moderating effects of resilience on depression in individuals with a history of childhood abuse or trauma exposure. Journal of Affective Disorders, 126(3), 411–414. 10.1016/j.jad.2010.04.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. World Health Organization. (2017). Depression: Fact Sheet. WHO. [Google Scholar]
  69. Zalewski M, Cyranowski JM, Cheng Y, & Swartz HA (2013). Role of maternal childhood trauma on parenting among depressed mothers of psychiatrically ill children. Depression and Anxiety, 30(9), 792–799. 10.1002/da.22116 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES